chat-quality-agent  by tanviet12

AI-powered customer service quality analysis agent

Created 2 weeks ago

New!

271 stars

Top 94.9% on SourcePulse

GitHubView on GitHub
Project Summary

Chat Quality Agent (CQA) is an AI-powered system designed to analyze and improve customer service quality. It automatically synchronizes messages from Zalo OA and Facebook Messenger, leverages AI models like Claude or Gemini to evaluate agent performance with detailed feedback and scoring, and dispatches automated alerts via Telegram or Email. This tool is beneficial for businesses seeking to gain objective insights into their customer support interactions and proactively address quality issues.

How It Works

The system operates by integrating with Zalo OA and Facebook Messenger to fetch customer conversations. It then employs AI models (Claude or Gemini) to assess the quality of customer support interactions, providing a pass/fail status, a score from 0-100, and detailed textual feedback. CQA can also categorize chats by custom topics like complaints or inquiries. An automated notification system sends alerts for critical issues, while a dashboard visualizes performance metrics and recent alerts. A "Batch AI mode" is available to optimize AI processing costs.

Quick Start & Requirements

  • Primary Install: The recommended installation method is an automated script: curl -s https://raw.githubusercontent.com/tanviet12/chat-quality-agent/main/install.sh | sudo bash. This script handles Docker setup, secret generation, image pulling, and service startup. Alternatively, build from source by cloning the repository, copying .env.example to .env, configuring environment variables, and running docker compose up -d --build.
  • Access: The application is accessible via http://your-server-ip or http://localhost. An initial setup page guides the creation of an admin account.
  • Prerequisites: Docker, Go 1.25+ (for building backend from source), MySQL 8.0, and API keys for Claude or Gemini are required.
  • SSL: Optional SSL can be configured by setting LEGO_DOMAIN and LEGO_EMAIL in .env and restarting the Nginx service. This requires a DNS A record pointing to the server IP.
  • Links:

Highlighted Details

  • AI-driven quality assessment using Claude or Gemini.
  • Direct message synchronization from Zalo OA and Facebook Messenger.
  • Automated alerts via Telegram and Email for immediate issue notification.
  • Multi-tenant architecture supporting multiple companies within a single instance.
  • Comprehensive dashboard featuring charts, statistics, and recent alerts.
  • Cost-saving "Batch AI mode" for processing multiple conversations efficiently.

Maintenance & Community

Detailed documentation is available at https://tanviet12.github.io/chat-quality-agent/, and a changelog is maintained in CHANGELOG.md. No specific community channels (like Discord or Slack) or information on major contributors, sponsorships, or partnerships are provided in the README.

Licensing & Compatibility

The project is licensed under the MIT License (SePay). This license is permissive and generally allows for commercial use and integration into closed-source projects.

Limitations & Caveats

Users must obtain and configure API keys for either Claude or Gemini to utilize the AI assessment features. Building from source requires Go 1.25+. While SSL is optional, its automatic setup necessitates DNS configuration. The README does not specify any known bugs, alpha/beta status, or unsupported platforms.

Health Check
Last Commit

1 week ago

Responsiveness

Inactive

Pull Requests (30d)
12
Issues (30d)
36
Star History
271 stars in the last 19 days

Explore Similar Projects

Feedback? Help us improve.